Blind Denoiser-based Beamspace Channel Estimation with Deep Learning in Millimeter-Wave Systems

In this paper, the problem of channel estimation for a lens-based array beamspace massive multiple-input multiple-output (MIMO) systems is investigated. Beamspace MIMO can utilize lens arrays to significantly reduce the number of radio frequency (RF) chains, however, the channel estimation is challenging due to a limited number of RF chains. To address this problem, we propose a novel blind denoiser combing CNN, GAN and DRN (CGDBD) for channel estimation in millimeter-wave massive MIMO systems. Numerical results show that the CGDBD network has satisfactory accuracy and a higher achievable sum rate (ASR). In addition, the CGDBD network can also remove unknown noise.